Correlates of human immunodeficiency virus infection in intravenous drug users: are treatment-program samples misleading?

Using a community-based sample of 1632 intravenous drug users in Baltimore, Maryland, USA, this study evaluated a possible selection bias faced when studying risk factors for human immunodeficiency virus, type 1 (HIV-1) in intravenous drug users drawn solely from treatment programs. Consenting subjects, recruited by extensive distribution of brochures and word-of-mouth, underwent confidential interviews about drug use behaviors in a setting that was independent of community service agencies. Of 1632 subjects, 275 (17%) were currently enrolled in treatment programs. These subjects were similar in many respects to those not currently in treatment, including HIV-1 seroprevalence; however, they differed in that those not in treatment were more likely to be male, younger and black than those currently in treatment, and to have started injecting drugs more recently. However, associations between HIV-1 seropositivity and a series of demographic and drug-using characteristics were mostly similar in direction and magnitude among subjects currently in treatment and those not in treatment. These results suggest that one-time samples of drug users enrolled in treatment programs do not necessarily present a misleading picture of correlates of HIV-1 seropositivity.